Ingham et al. BMC Genomics 2014, 15:1018http://www.biomedcentral.com/1471-2164/15/1018

RESEARCH ARTICLE Open Access

Dissecting the organ specificity of insecticideresistance candidate genes in Anopheles gambiae:known and novel candidate genesVictoria A Ingham1,2*, Christopher M Jones1, Patricia Pignatelli1, Vasileia Balabanidou3, John Vontas3,Simon C Wagstaff1, Jonathan D Moore2 and Hilary Ranson1*

Abstract

Background: The elevated expression of enzymes with insecticide metabolism activity can lead to high levels ofinsecticide resistance in the malaria vector, Anopheles gambiae. In this study, adult female mosquitoes from aninsecticide susceptible and resistant strain were dissected into four different body parts. RNA from each of thesesamples was used in microarray analysis to determine the enrichment patterns of the key detoxification genefamilies within the mosquito and to identify additional candidate insecticide resistance genes that may have beenoverlooked in previous experiments on whole organisms.

Results: A general enrichment in the transcription of genes from the four major detoxification gene families(carboxylesterases, glutathione transferases, UDP glucornyltransferases and cytochrome P450s) was observed in themidgut and malpighian tubules. Yet the subset of P450 genes that have previously been implicated in insecticideresistance in An gambiae, show a surprisingly varied profile of tissue enrichment, confirmed by qPCR and, for threecandidates, by immunostaining. A stringent selection process was used to define a list of 105 genes that aresignificantly (p ≤0.001) over expressed in body parts from the resistant versus susceptible strain. Over half of these,including all the cytochrome P450s on this list, were identified in previous whole organism comparisons betweenthe strains, but several new candidates were detected, notably from comparisons of the transcriptomes fromdissected abdomen integuments.

Conclusions: The use of RNA extracted from the whole organism to identify candidate insecticide resistance geneshas a risk of missing candidates if key genes responsible for the phenotype have restricted expression within thebody and/or are over expression only in certain tissues. However, as transcription of genes implicated in metabolicresistance to insecticides is not enriched in any one single organ, comparison of the transcriptome of individualdissected body parts cannot be recommended as a preferred means to identify new candidate insecticide resistantgenes. Instead the rich data set on in vivo sites of transcription should be consulted when designing follow upqPCR validation steps, or for screening known candidates in field populations.

Keywords: Detoxification, Insecticide resistance, Microarray, Mosquito, Transcriptome

* Correspondence: v.ingham@warwick.ac.uk; Hilary.Ranson@lstmed.ac.uk1Department of Vector Biology, Liverpool School of Tropical Medicine,Pembroke Place, Liverpool L35QA, UKFull list of author information is available at the end of the article

© 2014 Ingham et al.; licensee BioMed Central Ltd. This is an Open Access article distributed under the terms of the CreativeCommons Attribution License (http://creativecommons.org/licenses/by/4.0), which permits unrestricted use, distribution, andreproduction in any medium, provided the original work is properly credited. The Creative Commons Public DomainDedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article,unless otherwise stated.

Ingham et al. BMC Genomics 2014, 15:1018 Page 2 of 9http://www.biomedcentral.com/1471-2164/15/1018

BackgroundInsecticides play a vital role in reducing malaria transmissionin Africa. An escalation in the use of two very effective tools,indoor residual spraying with insecticides and insecticidetreated bednets, has led to impressive reductions in malariawith child death rates halved and more than 3.3 million livessaved since 2000 [1]. Inevitably, as insecticide use has inten-sified, malaria vectors have developed resistance to thesechemicals [2-4]. With just four classes of insecticides avail-able for public health and only the pyrethroids approved forbednet treatment, this poses a major challenge to sustainingand extending recent achievements in malaria reduction.Numerous studies have attempted to identify the genes

responsible for insecticide resistance in the major malariavectors. One of the most potent mechanisms identified todate is increased activity of enzymes that detoxify insecti-cides [4-6]. Four enzyme families are known to be associ-ated with insecticide metabolism (carboxylesterases (CCEs),glutathione transferases (GSTs), UDP glucornyltransferases(UGTs) and cytochrome P450s (P450s)) and a number ofindividual enzymes, most notably from the cytochromeP450 family, have been implicated in conferring resistanceto one or more insecticide classes [2,7-9]. More recently,the importance of interactions between different enzymesand transporters in the insecticide detoxification pathwayhas been recognised [10,11]. To dissect these pathways fur-ther, and to distinguish members of these large gene fam-ilies with housekeeping functions from those more likely tohave detoxification roles, further information on their sitesof expression is required. The first objective of this studywas to characterise expression patterns of the key genefamilies associated with insecticide resistance across themajor organs linked to xenobiotic detoxification in insects,with a particular focus on the P450s. Although transcrip-tomes of many of the key tissues in Anopheles gambiaehave already been described [12-15], this study used mater-ial from a highly insecticide resistant strain and from a sus-ceptible strain to identify genes whose tissue-specificenrichment might be linked to the resistance phenotype.To date, all comparisons of the transcriptome between

insecticide resistant and susceptible malaria vectors havecompared gene expression in the whole organism. Thisapproach has the potential to miss candidates. If, for ex-ample, expression of a gene is restricted to an organ thatcontributes only a small proportion of mRNA to the totalRNA pool, or differential expression occurs in only onetissue, even large differences in expression between a re-sistant and susceptible population may not be detectable[16]. Thus the second objective of the study was to com-pare gene expression in key body parts between an insecti-cide resistant and susceptible strain of mosquito toidentify candidates not immediately apparent in whole or-ganism microarray studies. Adult mosquitoes were dis-sected into body parts that could be readily separated with

minimal risk of contamination, and are suspected to be in-volved in metabolic resistance.

Results and discussionRNA was extracted from three dissected ‘body parts’: themalpighian tubules, the midgut and the abdomen in-tegument (containing the fat body, but also epidermal,neuronal, muscle and oenocyte cells) with the remainingundissected body parts forming a fourth sample group.Each biological replicate consisted of 15–20 adult femalemosquitoes from the major African malaria vector, An.gambiae. Dissections were performed on an insecticidesusceptible strain (N’Gousso) originally from Cameroon[17] and the Tiassalé strain from Cote d’Ivoire, which ishighly resistant to all four classes of insecticide [2,3].Transcription in these body parts was compared in two

ways (i) each body part against the corresponding wholeorganism for both strains and (ii) resistant against thecorresponding susceptible body parts (Additional file 1:Figure S1).

Gene enrichment in individual body part versus wholeorganismTranscripts showing enriched transcription in each ofthe four body parts were determined using a multipletest correction significance cut-off of p ≤0.05 for boththe resistant and susceptible strains. As expected, a clearpositive correlation can be observed for local transcrip-tion between the resistant and susceptible strains in eachbody part. This is illustrated in Figure 1, where the vastmajority of probes follow a y = x trend.In all body parts, similar numbers of transcripts showed

enriched or depleted transcription compared to the wholeorganism, with a range of 1.4% to 4.4% of the total probes(Additional file 2: Table S1). The magnitude of change intranscription of individual transcripts in the midgutand remaining body parts is relatively low with no tran-scripts exceeding log2 fold change of 2, compared to theabdomen integument and malpighian tubules where 7 and42 genes respectively are above log2 3-fold enriched(Figure 1). The full lists of transcripts enriched in eachbody parts are listed in Additional file 3: Table S2.The An. gambiae genome contains 211 genes encoding

four major detoxification families (111 cytochromeP450s, 31 GSTs, 43 CCES (only putatively catalyticallyactive enzymes included) and 26 UGTs), together com-prising 1.5% of the probes on the Agilent array [6]. Eachof the dissected body parts have an over representationof detoxification transcripts with more members of thesegene families overtranscribed relative to the whole or-ganism than found in the ‘remaining’ (undissected bodypart). Additional file 2: Table S1 shows that 8.5% of thetotal number of detoxification transcripts are enrichedin the malpighian tubules of the resistant strain, 13.7%

Figure 1 Gene enrichment in individual body parts in insecticide susceptible and resistant mosquitoes. Log2 fold change in transcriptionof all probes in individual tissues against the reconstituted whole organism, plotted for Tiassalé (resistant, x) and N’Gousso (susceptible, y). Probesfor the four detoxification families are indicated in pink (cytochrome p450s, 393 probes), green (carboxylesterases, 168 probes), red (GSTs, 152probes) and yellow (UDPs, 26 probes).

Ingham et al. BMC Genomics 2014, 15:1018 Page 3 of 9http://www.biomedcentral.com/1471-2164/15/1018

in the midgut and 7.6% in the abdomen integument. Incontrast, the ‘remaining’ (undissected body part) show adepletion of detoxification transcripts in the overtran-scribed subset when compared to the whole organism.These data reinforce the importance of the selected bodyparts in xenobiotic detoxification.The cytochrome P450 family has been most strongly

linked with insecticide resistance in Anopheles mosquitoes,with several enzymes capable of detoxifying insecticidesfrom more than one class [2,8]. Identifying the primary sitesof transcription of this enzyme family will aid prediction offunction [18] and help identify the key organs largely re-sponsible for insecticide detoxification in resistant mosqui-toes. Within the P450 gene family, body part enrichmentshows some relationship with the gene tree clustering(Additional file 4: Figure S2) with the CYP9J and 6P fam-ilies largely enriched in the midgut, the CYP4Gs, 6Ys and325Cs enriched in the abdomen integument and theCYP6Z family enriched in the malpighian tubules. The di-versity in enrichment patterns within this gene family ledus to look specifically at thirteen cytochrome P450s thathad been implicated in insecticide resistance to see if thesewere enriched in a particular body part (Figure 2). The cri-teria for inclusion of these P450s as candidate insecticideresistant genes was that they had been found to be signifi-cantly over expressed in pyrethroid and/or DDT resistantAn. gambiae populations in more than one independent

study. Four of these transcripts were significantly (p < 0.05)and highly (Log2 Fold Change >1.5) enriched in the malpig-hian tubules (CYP6M3, CYP6Z1, CYP6Z2 and CYP6Z3)compared to whole organism and one was enriched in themidgut and two in the abdomen integument (CYP4H24and CYP4G16/CYP4G17 respectively). The remaining 5showed no significant tissue enrichment.The enrichment of a subset of seven cytochrome P450s

in particular body parts was confirmed by qPCR, althoughfor two of these midgut enriched transcripts, a muchgreater over expression was observed by qPCR than formicroarray (Additional file 5: Figure S3) (A level of dis-cordance between qPCR and microarray is frequently ob-served [19] (but it in the current study, the direction ofchange was consistent between the two methodologies forall seven genes). Antibodies were available for three cyto-chrome P450s, which were used to confirm the major sitesof transcription within the abdomen integument of theTiassalé resistant strain. In agreement with microarraytranscription data, CYP6Z1 and CYP6Z2 were detected inthe malpighian tubules of resistant mosquitoes, whilstCYP4G17, identified as enriched in the abdomen integu-ment by both microarray and qPCR, was found only inthe oenocytes, a cellular layer located under the abdomencuticle (Figure 3). A CYP4G from Drosophila melanoga-ster, CYP4G1, has also been shown to be highly enrichedin oenocyte cells; this Drosophila enzyme catalyses a key

Figure 2 Local expression of cytochrome P450s linked toinsecticide resistance. Heatmap showing the log2 fold change of asubset of cytochrome p450s (implicated in insecticide resistance inprevious studies) in different body parts of the Tiassalé strain.Crosses indicate non-significance (p >0.05).

Ingham et al. BMC Genomics 2014, 15:1018 Page 4 of 9http://www.biomedcentral.com/1471-2164/15/1018

step in the formation of cuticular hydrocarbons [20]. Thepotential role of over expression of 4G17, and its paralo-gue 4G16, in altering the cuticular structure in insecticideresistant mosquitoes is currently under investigation.

A new list of candidate insecticide resistance genes fromthe body part specific arraysThe two strains used in this study originate from sites sepa-rated by approximately 2,500 km, so substantial variation ingene transcription between them is to be expected, regard-less of the difference in their insecticide resistance profile.Thus caution must be applied when correlating gene tran-scription levels with the resistance phenotype. However,with a goal of identifying further candidate insecticide resist-ance genes from the direct comparisons between dissectedbody parts from the resistant and susceptible strain for fur-ther functional validation, we applied a stringent selectionprocess to derive a new gene list. This involved selectingonly genes detected as enriched in body parts from Tiassaléversus N’Gousso via both the limma and GaGa methodswith an adjusted p value ≤0.001 and setting a cut off of 1.4-fold differential expression (see Methods section). Thisidentified a list of 134 transcripts, representing 105 genestranscribed at higher levels in the Tiassalé strain and 16genes with higher transcription in the susceptible N’Goussostrain (Additional file 6: Table S3). Eleven of these tran-scripts were selected for qPCR validation yielding a positivecorrelation with the array (Pearson’s correlation (r = 0.870))(Figure 4).All of the 9 cytochrome P450s (CYP6P2, P3 and P4,

CYP6Z2, Z3 CYP4H17, CYP4K2, CYP4C35 and CYP4G16)on this candidate list were also detected in the directwhole comparisons of the whole organism transcriptomesbetween these strains [3 and C. Strode, unpublished data].Indeed the majority (89) of the 134 transcripts on our can-didate list derived from the comparisons of dissected bodyparts were also detected in comparisons of the entire tran-scriptome between the two strains. Nevertheless, 22(16.4%) would have been missed using whole organismarrays and 23 (17.2%) of the genes are regulated in the op-posite direction between the whole organism and dis-sected body part comparisons. Of the 22 transcripts notdetected in strain comparisons at the whole organismlevel, the majority (13) were detected from the abdomenintegument (Table 1). Abdomen integuments from otherresistant strains of An. gambiae are being dissected tosearch for further supporting evidence for a role of thesetranscripts in conferring resistance prior to follow upfunctional analysis.

ConclusionsMicroarrays are widely used to identify insecticide resist-ance mechanisms in mosquito populations [3,9,21,22].However, all of these studies have used whole-organism

CYP4G17 Ovaries

Malp.tubul.

Oe

Oe

Malp..tubul

Gut

CYP6Z2

DNA MERGE

Malp.tubul.

Ovaries

CYP6Z1A

B

C

Ovaries

Figure 3 Immunohistochemical stainings of cytochrome P450s associated with pyrethroid resistance. Longitudinal sections from resistantmosquito (Tiassalé) specimens immunostained with A) a-CYP6Z1, B) a-CYP6Z2 and C) a-CYP4G17 specific antibodies (left panel, green color). Themiddle panel shows the same sections stained red using the nucleic acid stain TOPRO. The merged immunohistochemical staining (P450 stainingsand nuclei) appears in the right column. Bar scale (yellow line): 100 nm.

Ingham et al. BMC Genomics 2014, 15:1018 Page 5 of 9http://www.biomedcentral.com/1471-2164/15/1018

RNA as a template, which may dilute or distort the finallevel of transcript expression detectable. By dissecting someof the major body parts involved in xenobiotic detoxificationfrom two strains of An. gambiae, differing in their resistancephenotype we have been able to simultaneously identify the

Figure 4 Validation of selection of genes from the candidate list by qcomparison between the resistant and susceptible populations [29,30]. Genprobes enriched in the resistant versus susceptible strain in one or more d

local expression profiles of known insecticide resistance can-didates and compare transcription between the two strainswithin individual body parts.This rich data set will be useful for establishing path-

ways of detoxification as genes catalysing the three classic

PCR. Log2 fold change from qPCR data and array data of the directes were selected from the ‘stringent candidate gene list’ consisting ofissected body parts. Standard error bars are shown.

Table 1 Genes showing the greatest differential expression between resistant and susceptible strains in specific bodyparts, but which were not identified in whole transcriptome comparisons

AGAP identifier Description Tissue Log2 fold change

AGAP006710 Chymotrypsin-1 Midgut 7.20

AGAP006741 Putative Tight Junction Associated Protein Abdomen Integument 2.33

AGAP001769 Beat Protein Abdomen Integument 2.27

AGAP007650 GADD45 (Growth Arrest and DNA Damage Inducible Protein) Abdomen Integument 2.15

AGAP000717 Monocarboxylate Transporter Midgut 1.52

AGAP005563 Sugar Transporter Abdomen Integument 1.91

AGAP009521 Ankyrin Erythrocytic Abdomen Integument 1.86

AGAP007879 Steroid Dehydrogenase Midgut 1.74

AGAP013219 Elongase Abdomen Integument 1.70

AGAP007301 Conserved hypothetical protein Malpighian Tubules and Midgut 1.61

AGAP001649 Glucosylceramidase Abdomen Integument 1.11

AGAP006358 TTC27 Abdomen Integument 1.05

AGAP002517 Unknown Abdomen Integument 0.95

AGAP007691 Serpin 18 Abdomen Integument 0.92(−RC)

0.87(−RB)

AGAP005651 Cytoplasmic tRNA 2-thiolation protein Abdomen Integument 0.80

AGAP004099 Conserved Hypothetical Protein Abdomen Integument 0.60

AGAP005334 C-Type Lectin Midgut −3.62

AGAP003271 Anexin B10B Midgut −1.87

AGAP002752 DNAJ Homolog Remaining Body Parts −1.10

Splice variants, as predicted by VectorBase, are labelled with variations of -Rx.

Ingham et al. BMC Genomics 2014, 15:1018 Page 6 of 9http://www.biomedcentral.com/1471-2164/15/1018

classes of drug metabolism (oxidation, conjugation andexcretion) [23] would be expected to be co-regulated. Pyr-ethroid mimetic activity–based probes, used to detect pyr-ethroid metabolising enzymes in the rat liver, identified apotential network of drug metabolising enzymes from mul-tiple families involved in pyrethroid metabolism [10]. Apply-ing the same approach to insecticide resistant mosquitoesand using the data on local transcription from the currentdata set, will help unravel the pathways of insecticide metab-olism selected for by intensive use of pyrethroids.Although potential new insecticide resistance candi-

dates have emerged from this study, it is encouragingthat the majority of candidate insecticide resistant tran-scripts identified from direct comparison of the tran-scriptomes of dissected body part were also detected inthe whole organisms comparisons. No single body partemerged as the key site of overtranscription of putativeinsecticide resistance genes in this study and it is there-fore recommended that, unless resources enable a morecomprehensive study design involving multiple dissectedtissues, transcriptional approaches to identify candidateinsecticide resistance transcripts continue to use thewhole body transcriptome. Nevertheless this data set onlocal sites of transcription should be consulted when de-signing follow up qPCR validation steps, or for screeningknown candidates in field populations.

MethodsMosquito rearing conditionsThe An. gambiae used in these experiments were allreared under standard insectary conditions at 27°C and70-80% humidity under a 16:8 hour photoperiod. TheN’Gousso strain is originally from Cameroon and is sus-ceptible to all classes of insecticide [17]. N’Guosso is theM molecular form of A gambiae, recently re-classified as aseparate species, Ano coluzzi [24]. In contrast, the Tiassaléstrain from Côte D’Ivoire is resistant to all classes of in-secticide [2,3]. This strain was colonised from the fieldsite in 2012 and is a mixture of the M and S molecularforms. At the time of the study, the LD50 for the Tiassaléstrain was 68 and 81 fold higher than the correspondingvalue for the N’Gousso strain for permethrin and delta-methrin respectively. Further details of the resistanceprofile of this strain are contained within references byEdi et al. [2,3].

Microarray experimentsRNA was extracted from three dissected body parts: themalpighian tubules, the midgut and the abdomen in-tegument (containing the fat body but also epidermal,neuronal, muscle and oenocyte cells) with the remainingundissected body parts forming a fourth sample group.Mosquitoes were collected between the hours of 8AM

Ingham et al. BMC Genomics 2014, 15:1018 Page 7 of 9http://www.biomedcentral.com/1471-2164/15/1018

and 2 pm and dissected immediately on a CO2 block. Postdissection, each body part was added to extraction bufferfrom the PicoPure RNA extraction kit, heated for 30 mi-nutes at 42°C and frozen at −80°C as per manufacturersinstructions. Each biological replicate for each strainconsisted of RNA, extracted using PicoPure RNA Isola-tion kit (Arcturus), from 12 3–5 day old non-blood fed,presumed mated females. The quantity and quality ofthe RNA was assessed using a nanodrop spectropho-tometer (Nanodrop Technologies UK) and Bioanalyser(Agilent) respectively. Four biological replicates wereprepared for each body parts per strain. RNA from thefour dissections was pooled according to the proportionof RNA extracted from each body part to reconstitutethe ‘whole organism’ (7%, 6%, 24% and 63% RNA fromabdomen integument, malpighian tubules, midgut, andremaining material respectively). The use of a reconsti-tuted reference sample minimised potential sources ofbias that could have arisen from circadian changes ingene expression and changes in the proportion of the Mor S molecular form in the different biological repli-cates. 100 ng of RNA was amplified and labelled withCy3 and Cy5, using the Two colour low input QuickAmp labelling kit (Agilent) following the manufacturersinstructions. Samples were then purified (Qiagen) withthe cRNA yield and quality assessed using the nano-drop and Bioanalyser respectively. RNA from each Tias-salé body part was competitively hybridised with the re-spective N’Gousso body part, as well as each body partfrom the resistant and susceptible strain being com-pared to the re-constituted whole organism (Additionalfile 1: Figure S1). Dye swaps were performed on two outof four technical replicates for each array, to correct fordye bias.Labelled cRNAs were hybridised to the whole genome

8×15k An gambiae array (ArrayExpress accession numberA-MEXP-2211). Microarray hybridisation, washing andscanning were performed according to previously de-scribed protocols [8].

Microarray analysisThe resulting data were analysed using R. Within-arraynormalisation was carried out by Loess, and between arraynormalisation by Aquantile. Signals were corrected for dyeautomatically. The limma package [25] was used to fit lin-ear models to the normalised data. In the case of completeloop designs a design matrix was used to infer the appropri-ate contrast matrices for each array. All parameters usedwere default. A bespoke pipeline using the GaGa package[26] was used to fit gamma-gamma models of variation tonormalised corrected signals, in order to assign probes toone of two patterns of expression X equals Y or X does notequal Y, where X represents the resistant population arraysand Y the susceptible arrays. These data were subsequently

used to assess enrichment in each expression pattern,through GO term analysis using the TopGO package [27].A standard FDR adjusted p value cut off of p ≤0.05 was ap-plied to all data describing localisation of detoxificationcandidates. A second stringent selection method was usedto reduce the probe list based on previously publishedmethodology [19], requiring that the following criteria weremet: adjusted p-value ≤0.001, raw fluorescence intensity >median, and Tiassalé vs. N’Gousso ±0.485 Log2 fold changebetween the strains. All candidates selected also demon-strated a positive GaGa analysis fold change, indicative ofhigher transcript localisation in the resistant tissue to thesusceptible, thereby utilising all available array data.

RT-qPCRRNA (4 μg) from each biological replicate was reversetranscribed using Oligo dT (Invitrogen) and SuperscriptIII (Invitrogen) according to manufacturers instructions.Quantitative real-time PCR was performed using SYBRGreen Supermix III (Applied Biosystems) using anMX3005 and the associated MxPro software (Agilent).Primer Blast (NCBI) [28] was used to design primer pairs(Additional file 7: Table S4). Where possible, primers weredesigned to span an exon junction but this was not pos-sible for six of the P450 genes (CYP325A1, CYP6P3,CYP4G17, CYP6Z3, CYP12F2 and CYP6Z2) due to thehigh degree of polymorphisms in their DNA sequence.Each 20 μl reaction contained 10 μl SYBR Green Super-mix, 0.3 μM of each primer and 1 μl of 1:10 dilutedcDNA. Standard curves were produced using wholeN’Gousso cDNA, in 1, 1:5, 1:25, 1:125 dilutions, (48.3 ng/μl to 0.386 ng/μl). qPCR was performed with the followingconditions: 3 minutes at 95°C, with 40 cycles of 10 secondsat 95°C and 10 seconds at 60°C. All amplification efficien-cies of designed primers were within acceptable range (90-120%), following MIQE guidelines [29].

Preparation of antibodiesFragments encoding unique peptides for CYP6Z1 andCYP4G17 were cloned into the pET 16b vector. Uponexpression, the resulting His-tagged peptide was purifiedto homogeneity by Ni-NTA affinity chromatography andused to raise rabbit polyclonal antibodies. The CYP6Z1peptide sequence was: VALRDLNNPDSFINNIRTAGVFLCPGLLKFTGINSLSPPMKKFTTEVISSHLHQRETGQVTRKDFIQMLTDLRRKAGSSGEETLTDA and the CYP4G17 peptide: KRQLKIHLRLDPLFNLTGVKKEQERLLQIIHGLTRKVVREKKQLYERQMAEGKMPSPSLTEIIGKEEKPGEGQLGGSPAFISQ. The antibody for CYP6Z2 was a giftfrom Dr Mark Paine (LSTM, UK). Rabbit antibodies toCYP6Z2 were prepared to the C terminal peptide se-quence MRIDHRK by Moravian Biotechnology, Brno,Czech Republic.

Ingham et al. BMC Genomics 2014, 15:1018 Page 8 of 9http://www.biomedcentral.com/1471-2164/15/1018

Immunofluorescence and microscopyFemale mosquitoes (3–5 days old) were fixed in cold solu-tion of 4% PFA (methanol free, Thermo scientific) inphosphate-buffered saline (PBS) for 4 h and then werecryo-protected in 30% sucrose/PBS at 4°C for 12 h. Finally,mosquitoes were immobilized in O.C.T. (Tissue-Tek,SAKURA) and stored at -80°C until use.Immunofluorescent analysis, followed by confocal mi-

croscopy, was performed to longitudinal sections of frozenpre-fixed mosquito specimens. More detailed, 10 mm sec-tions, obtained in Leitz kryostat 1720 digital, were washed(3 × 5 min) with 0,05% Tween in PBS and blocked for 3 hin blocking solution (1% Fetal Bovine Serum, biosera, in0,05% Triton/PBS). Then, the sections were stained withrabbit primary antibodies in 1/500 dilution, followed bygoat anti-rabbit (Alexa Fluor 488, Molecular Probes)(1/1000) that gave the green colour. Also To-PRO 3-Iodide(Molecular Probes), which stains DNA specifically (redcolour), was used, after RNAse A treatment. As controls,pre-immune serums (in 1/500 dilutions) and anti-rabbit(Alexa Fluor 488, 1/1000) were tested, in parallel with a-P450’s to check specificity of each primary antibody. Finallyimages were obtained on Leica TCS-NT Laser Scanningmicroscope using the 40-objective.

Supporting informationThe data sets supporting the results of this article are in-cluded within the article and its additional files. All micro-array datasets are MIAME compliant [30] and deposited inArrayExpress and VectorBase (Accession numbers E-MTAB-2808).

Additional files

Additional file 1: Figure S1. Schematic of design of microarrayexperiment. a. Sample vs reference design, for each of the susceptible labpopulation N’Gousso and the resistant population Tiassalé b. Resistant (T)vs Susceptible (N) design, for each individual tissue.

Additional file 2: Table S1. Overview of probes over or undertranscribed in each body part for both the resistant and susceptiblestrains when compared to the whole organism. The local transcription ofprobes in each of the two mosquito strains is expressed as total numberof probes over or under transcribed in a particular body part, comparedto the reconstituted whole organism, as a percentage of the total probeson the array. For genes that were represented by multiple probes, anaverage of all the probes was used.

Additional file 3: Table S2. Genes significantly enriched or depleted inone or more body part compared to the whole organism. AGAPidentifiers, description and the body part in which the gene is enrichedwith log2 fold change (relative to reconstituted whole transcriptome)indicated for each strain. Only probes outside the 95% intervals are listed.Blue indicates cytochrome p450s, green GSTs, purple COEs, orange UDPsand teal carboxylesterases.

Additional file 4: Table S3. Probe list from stringent analysis of directtranscriptome comparisons of dissected body parts from susceptible andresistant strains. AGAP identifier, description, body part that the probe fitsthe stringent selection criteria, whole organism array log2 fold change,resistant vs susceptible array log2 fold change, GaGa log2 fold change

and log2 body part qPCR results for given tissue. qPCR validation hasbeen performed on several candidates. Cells are coloured with a gradientdependent upon the directionality of the fold change, down regulatedtranscripts are indicated in red and up regulated in green. Sheets forboth up regulated and down regulated genes are present.

Additional file 5: Figure S2. Local expression of all Cytochrome p450s,following a phylogenetic dendrogam. Full length protein sequencealignment and neighbour joining tree as computed on MEGA5decorated with local log2 transcription of Tiassalé cytochrome p450s.

Additional file 6: Figure S3. qPCR validation of body part enrichmentof cytochrome P450s. Difference in transcription levels in Tiassalé RNAfrom individual dissected body parts, compared to the reconstitutedwhole, were measured by qPCR for six P450 genes and compared to thedata obtained from the microarray. Data represent log2 fold change withassociated standard error bars.

Additional file 7: Table S4. qPCR primer list. Forward and reverseprimers used for all qPCR reactions. All primer products are between 80and 150 base pairs and follow MIQE guidelines.

AbbreviationscDNA: Complementary DNA; cRNA: Complementary RNA; FDR: Falsediscovery rate; GO: Gene ontology; mRNA: Messenger RNA;P450s: Cytochrome p450s; PBS: Phosphate-buffered saline; qPCR: QuantitativePCR; UDP: Uridine disphosphate.

Competing interestsThe authors declare that they have no competing interests.

Authors’ contributionsV.I, C.M.J. P.P and V.B. performed the experiments. H.R and J.V, designed thestudy. V.I, S.W and J.M. analysed the data. V.I and H.R drafted the manuscript.All authors read and approved the final manuscript.

AcknowledgementsThe research leading to these results was supported by the European UnionSeventh Framework Programme FP7 (2007–2013) under grant agreement no265660 AvecNet. We thank Dr Clare Strode, LSTM, for sharing unpublishedmicroarray data and Dr Gareth Lycett, LSTM, for critical review of themanuscript.

Author details1Department of Vector Biology, Liverpool School of Tropical Medicine,Pembroke Place, Liverpool L35QA, UK. 2Systems Biology DTC, University ofWarwick, Senate House, Coventry CV47AL, UK. 3Department of Biology,University of Crete, Vassilika Vouton, 71409 Heraklion, Greece.

Received: 25 July 2014 Accepted: 14 November 2014Published: 25 November 2014

References1. World Health Organisation: World Malaria Report. Geneva: 2013. http://www.

who.int/malaria/publications/world_malaria_report_2013/en/.2. Edi CV, Djogbénou L, Jenkins AM, Regna K, Muskavitch MAT, Poupardin R,

Jones CM, Essandoh J, Kétoh GK, Paine MJI: CYP6 P450 enzymes and ACE-1 duplication produce extreme and multiple insecticide resistance in themalaria mosquito anopheles gambiae. PLoS Genet 2014, 10:e1004236.

3. Edi CV, Benjamin KG, Jones CM, Weetman D, Ranson H: Multiple-insecticideresistance in anopheles gambiae mosquitoes, Southern Côte d’Ivoire.Emerg Infect Dis 2012, 18:1508–1511.

4. Ranson H, N’Guessan R, Lines J, Moiroux N, Nkuni Z, Corbel V: Pyrethroidresistance in African anopheline mosquitoes: what are the implicationsfor malaria control? Trends Parasitol 2011, 27:91–98.

5. Liu N, Xu Q, Zhu F, Zhang L: Pyrethroid resistance in mosquitoes.Insect Sci 2006, 13:159–166.

6. Ranson H, Claudianos C, Ortelli F, Abgrall C, Hemingway J, Sharakhova MV,Unger MF, Collins FH, Feyereisen R: Evolution of supergene familiesassociated with insecticide resistance. Science 2002, 298(80-):179–181.

Ingham et al. BMC Genomics 2014, 15:1018 Page 9 of 9http://www.biomedcentral.com/1471-2164/15/1018

7. Chiu T-L, Wen Z, Rupasinghe SG, Schuler MA: Comparative molecularmodeling of Anopheles gambiae CYP6Z1, a mosquito P450 capable ofmetabolizing DDT. Proc Natl Acad Sci 2008, 105:8855–8860.

8. Mitchell SN, Stevenson BJ, Müller P, Wilding CS, Egyir-Yawson A, Field SG,Hemingway J, Paine MJI, Ranson H, Donnelly MJ: Identification andvalidation of a gene causing cross-resistance between insecticide classesin anopheles gambiae from Ghana. Proc Natl Acad Sci 2012,109(16):6147–6152.

9. Müller P, Warr E, Stevenson BJ, Pignatelli PM, Morgan JC, Steven A, YawsonAE, Mitchell SN, Ranson H, Hemingway J, Paine MJI, Donnelly MJ: Field-caught permethrin-resistant anopheles gambiae overexpress CYP6P3, aP450 that metabolises pyrethroids. PLoS Genet 2008, 4:e1000286.

10. Ismail HM, O’Neill PM, Hong DW, Finn RD, Henderson CJ, Wright AT, CravattBF, Hemingway J, Paine MJI: Pyrethroid activity-based probes for profilingcytochrome P450 activities associated with insecticide interactions.Proc Natl Acad Sci 2013, 10(49):19766–19771.

11. Chandor-Proust A, Bibby J, Regent-Kloeckner M, Roux J, Guittard-Crilat E,Poupardin R, Asam Riaz M, Paine M, Dauphin-Villemant C, Reynaud S: Thecentral role of mosquito cytochrome P450 CYP6Zs in insecticidedetoxification revealed by functional expression and structuralmodelling. Biochem J 2013, 455(1):75–85.

12. Baker DA, Nolan T, Fischer B, Pinder A, Crisanti A, Russell S: Acomprehensive gene expression atlas of sex-and tissue-specificity in themalaria vector. Anopheles gambiae. BMC Genomics 2011, 12:296.

13. Koutsos AC, Blass C, Meister S, Schmidt S, MacCallum RM, Soares MB, CollinsFH, Benes V, Zdobnov E, Kafatos FC, Christophides GK: Life cycletranscriptome of the malaria mosquito Anopheles gambiae andcomparison with the fruitfly Drosophila melanogaster. Proc Natl Acad Sci2007, 104(27):11304–11309.

14. Marinotti O, Calvo E, Nguyen QK, Dissanayake S, Ribeiro JMC, James AA:Genome-wide analysis of gene expression in adult Anopheles gambiae.Insect Mol Biol 2006, 15:1–12.

15. Neira Oviedo M, Vanekeris L, Corena-Mcleod MDP, Linser PJ: A microarray-based analysis of transcriptional compartmentalization in the alimentarycanal of Anopheles gambiae (Diptera: Culicidae) larvae. Insect Mol Biol2008, 17:61–72.

16. Johnson BR, Atallah J, Plachetzki DC: The importance of tissue specificityfor RNA-seq: highlighting the errors of composite structure extractions.BMC Genomics 2013, 14:586.

17. Harris C, Lambrechts L, Rousset F, Abate L, Nsango SE, Fontenille D, MorlaisI, Cohuet A: Polymorphisms in Anopheles gambiae immune genesassociated with natural resistance to Plasmodium falciparum.PLoS Pathog 2010, 6:e1001112.

18. Chung H, Sztal T, Pasricha S, Sridhar M, Batterham P, Daborn PJ:Characterization of Drosophila melanogaster cytochrome P450 genes.Proc Natl Acad Sci 2009, 106:5731–5736.

19. Morey JS, Ryan JC, Van Dolah FM: Microarray validation: factorsinfluencing correlation between oligonucleotide microarrays and real-time PCR. Biol Proced Online 2006, 8:175–193.

20. Qiu Y, Tittiger C, Wicker-Thomas C, Le Goff G, Young S, Wajnberg E, FricauxT, Taquet N, Blomquist GJ, Feyereisen R: An insect-specific P450 oxidativedecarbonylase for cuticular hydrocarbon biosynthesis. Proc Natl Acad Sci2012, 109:14858–14863.

21. Jones CM, Toé HK, Sanou A, Namountougou M, Hughes A, Diabaté A,Dabiré R, Simard F, Ranson H: Additional Selection for InsecticideResistance in Urban Malaria Vectors: DDT Resistance in Anophelesarabiensis from Bobo-Dioulasso, Burkina Faso. PLoS One 2012, 7:e45995.

22. Tene BF, Poupardin R, Costantini C, Awono-Ambene P, Wondji CS, RansonH, Antonio-Nkondjio C: Resistance to DDT in an urban setting: commonmechanisms implicated in both M and S forms of anopheles gambiae inthe City of Yaoundé Cameroon. PLoS One 2013, 8:e61408.

23. Gibson GG, Skett P: Introduction to Drug Metabolism. Cheltenha, UK: NelsonThornes; 2001.

24. Coetzee M, Hunt RH, Wilkerson R, Della Torre A, Coulibaly MB, Besansky NJ:Anopheles coluzzii and Anopheles amharicus, new members of theAnopheles gambiae complex. Zootaxa 2013, 3619:246–274.

25. Smyth GK: Linear models and empirical bayes methods for assessingdifferential expression in microarray experiments. Stat Appl Genet Mol Biol2004, 3:3.

26. Rossell D: GaGa: a parsimonious and flexible model for differentialexpression analysis. Ann Appl Stat 2009, 3:1035–1051.

27. Alexa A, Rahnenfuhrer J: topGO: Enrichment Analysis for Gene Ontology. RPackage Version 2.18.0. ; 2010. http://www.bioconductor.org/packages/release/bioc/html/topGO.html.

28. Ye J, Coulouris G, Zaretskaya I, Cutcutache I, Rozen S, Madden TL: Primer-BLAST: a tool to design target-specific primers for polymerase chainreaction. BMC Bioinformatics 2012, 13:134.

29. Bustin SA, Benes V, Garson JA, Hellemans J, Huggett J, Kubista M, Mueller R,Nolan T, Pfaffl MW, Shipley GL: The MIQE guidelines: minimuminformation for publication of quantitative real-time PCR experiments.Clin Chem 2009, 55:611–622.

30. Brazma A, Hingamp P, Quackenbush J, Sherlock G, Spellman P, Stoeckert C,Aach J, Ansorge W, Ball CA, Causton HC: Minimum information about amicroarray experiment (MIAME)—toward standards for microarray data.Nat Genet 2001, 29:365–371.

doi:10.1186/1471-2164-15-1018Cite this article as: Ingham et al.: Dissecting the organ specificity ofinsecticide resistance candidate genes in Anopheles gambiae: knownand novel candidate genes. BMC Genomics 2014 15:1018.

Submit your next manuscript to BioMed Centraland take full advantage of:

• Convenient online submission

• Thorough peer review

• No space constraints or color figure charges

• Immediate publication on acceptance

• Inclusion in PubMed, CAS, Scopus and Google Scholar

• Research which is freely available for redistribution

Submit your manuscript at www.biomedcentral.com/submit